Unsupervised Learning Summarization Templates from Concise Summaries

Horacio Saggion

We here present and compare two unsupervised approaches for inducing the main conceptual information in rather stereotypical summaries in two different languages. We evaluate the two approaches in two different information extraction settings: monolingual and cross-lingual information extraction. The extraction systems are trained on auto-annotated summaries (containing the induced concepts) and evaluated on human-annotated documents. Extraction results are promising, being close in performance to those achieved when the system is trained on human-annotated summaries.